[Home] [Lectures] [Next]

Impacts of Severe Weather and Possible Trends in Severe Weather

Impacts of Extreme Events in The United States

[nws fatalities]
Severe weather fatalities in the United States. (Source: National Weather Service)

The figure to the left gives the average number of severe weather fatalities in the United States per year based on the 2006-2015 period (10 year average) and the 1986-2015 period (30 year average). The source of this information is the National Weather Service (NWS). The 10 year average total number of weather related fatalities per year in the United States is 543. Because there are many different ways to estimate the impacts of severe weather, if you check other sources, you will likely find different numbers. Severe weather can be deadly. Each year in the US, federal, state, and local governments spend a good deal of money in the areas of severe weather prediction, preparation and relief. For the most part this money spent has been effective in limiting the number of US deaths due to all types of severe weather. Considering all the severe weather events that are reported in the news and the number of people affected by these events throughout the US, the average number of deaths per year caused by these events is relatively low. Inevitably there will always be some deaths due to extreme weather, in spite of our best efforts to prevent it.

To put these numbers into perspective, the average number of United States deaths related to vehicle accidents has been over 30,000 each year since 1946 (see Wikipedia list of motor vehicle deaths in the US). In addition, in the United States an average of over 16,000 people have been murdered per year over the last decade (see US crime statistics 1960 - 2014). One point will be made with respect to statistics like this. If one of our goals as a society is to most effectively reduce the number of "preventable" deaths, it seems like we would focus more on reducing the relatively large numbers of those killed in vehicle accidents or murders, rather than spending so much money and effort to further reduce the relatively small number of people who die in severe weather events.

There are rather large changes in loss of life for each category of severe weather from year to year. This is shown in the National Weather Service's compilation of severe weather deaths from 1940 - 2015. If you look through the statistics, you will notice that the number of deaths due to lightning, tornadoes, and hurricanes have generally been lower in recent decades compared to the 1940 through 1980 period. This is most likely the result of better forecasts and overall preparedness for severe weather events. Deaths due to floods have not changed much. This may be because there are more people in flood-prone areas than there were in previous decades. The NWS began keeping track of deaths caused by exposure to heat and cold, wind, and rip currents more recently. Note that a significant portion of the deaths are associated with deadly sporatic events, such as the 1000+ deaths in the 1995 heatwave, the 1000+ deaths due to hurricane Katrina in 2005, and the 500+ deaths due tornado outbreaks in 2011.

Beside deaths, severe weather is the cause of extreme economic loss as well. The United States has sustained 196 severe weather disasters in which the overall cost has exceeded $1 billion dollars. Those events are listed and described in this Text summary US billion dollar weather disasters from 1980 - 2016

Trends in Severe Weather Deaths and Damages in the United States

Consider the following question:

If the answer to the above question is yes, then we might ask:

Keep in mind that there are many different ways to estimate the impacts of severe weather and that different studies will produce different results. Thus, conclusions drawn from a single study can be considered somewhat subjective. One problem with determining trends in the impacts of severe weather events is that there are large fluctuations from year to year in the number and severity of extreme weather events. For example consider hurricanes. Some years there are no major hurricanes that strike the United States, while in other years, multiple major hurricanes may strike. Exceptionally extreme events occur infrequently, perhaps once a decade or once a century. Thus, the choice of time period over which to analyze trends will affect the results. A second problem is that the choice of geographical area selected for study will affect the results, since extreme weather events are often localized. For example, we already pointed out that hurricane activity in the Atlantic has been high since 1995 (probably due to a natural 20-40 year cycle), but there has not been a world-wide increase in hurricane activity. This is also a problem when focussing on small regions of the United States, as extreme events may have increased in one state over the last decade, but decreased in another state.

As for death counts, there are no standard methods for determining which deaths should be attributed to a weather event, which leads to differences among studies. This is generally not a problem for things like hurricanes or tornadoes where cause of death is obviously attributable to the weather event, but what about something like a heat wave? If someone dies from a heart condition during the heat wave, was it because that person was stressed by the heat or was that person about to die anyway? Some studies will only count a heat death if the a coroner pronounces cause of death as heat stroke and not something like heart failure. This issue is further discussed in the last section of this page.

It is even more difficult to standarize the economic costs of severe weather events. We can argue about how to answer the following questions: What constitutes a "cost"? Do we want to look at insured losses or all losses? How can we be sure that loss estimates are accurate for individual disasters and/or that they are comparable across disasters? How can we possibly compare the relatively high property losses in developed countries with the relatively high social costs (such as deaths, injuries, and homelessness) in developing nations? Which indirect costs should be included?

Keeping these difficulties in mind, I will now provide a summary for trends in the US only based on my reading and interpretation of available information on trends in severe weather events and severe weather impacts. Please keep in mind that the information is surprisingly scarce and different sources are often contradictory.

It is difficult to determine if there have been any significant trends in loss of life due to extreme weather events in recent years (say over the last couple of decades or so). For the most part, it appears that the number of deaths due to exteme weather events has not changed significantly over the last 20 years. Prior to the most recent 20 years in the United States, there was a continuing drop in the number of deaths caused by severe weather events, from highest in the early part of the 20th century to lowest in that latter part of the century. This was most likely the result of improving weather forecasts, weather warnings, and public education and preparedness. Over the last 20 years or so, improvements in forecasting and warning are happening at a slower rate. Thus, the United States may have reached a balance between an increasing population, which exposes more people to possible weather-realted deaths, and continuing modest improvements in weather forecasting and public preparation, which tends to reduce the number of deaths due to extreme weather.

In recent years, most types of extreme weather events have led to larger economic losses in the United States (even after adjusting for inflation). Multi-billion dollar losses now occur with increasing frequency (see information provided in previous section). The trend in losses has led many to conclude that the United States has witnessed changes in the frequency and/or intensity of extreme events. However, the majority of evidence suggests that it is not changes in the frequency/severity of extreme weather events that are responsible for the increasing cost, but rather much of the increase in property damage probably reflects from the fact that property values and population are increasing. Thus, when severe weather does strike, it is more likely to impact structures (there are more of them) and the structures themselves are more highly valued (even after considering inflation). This is basically the same conclusion we reached about trends in hurricane damages.

Long Term Trends in Worldwide Extreme Weather Events?

There have been many recent news stories about extreme weather events and their effects on people both within the United States and around the world. A recurring question is whether or not these extreme weather events are becoming more frequent and severe in recent years and if so what might be the causing it. This question is briefly discussed in this section, however, a definitive answer is not given as there continues to be much scientific debate on this issue. I encourage all of you to keep an open mind on this issue, especially when considering the arguments of anyone who claims to know for sure.

The first question to address is can we determine if the number of extreme weather events is actually increasing or even changing significantly with time? In order to answer that we would need to have some idea about what is "normal" or "expected." By definition extreme events are rare. We also know from history that extreme weather events have always happened and that they are not evenly distributed in time, e.g., some past decades had more extreme events than others. Are these past fluctuations in extreme weather events randomly distributed or are there some underlying patterns in time? We know that there are natural cycles of climate oscillations that operate on a range of different time scales. For example, El Nino/La Nina oscillations happen on timescales of a few years; Hurricanes in the north Atlantic Ocean oscillate over multi-decadal timescales, which is something we covered earlier; Ice Age cycles have been occurring over timescales of about 100,000 years. All of these natural climate oscillations likely influence weather patterns and hence influence the liklihood of extreme weather events. We certainly do not understand all of these natural climate cycles that occur on Earth and it is likely that they interact with each other. In addition, given that some of these natural climate cycles have only recently been recognized, there are probably other cycles that we do not even know about. Our incomplete understanding of these cycles combined with the fact that good observations of extreme weather events have only been available for recent decades, makes it difficult for us to determine if there really are more extreme weather events today than there "should be." One must also keep in mind that there are more people and more weather observations than ever. Unnoticed extreme events in the past (perhaps because there were no people living where the event took place) can make it appear as if there are increasing numbers of extreme events in recent years, even when there may not be a real change in the number of extreme weather events.

Let's go though a hypothetical example that illustrates some of the potential problems with having only a limited understanding of the underlying statistics on extreme weather events. Assume that there were more extreme weather events in the first decade of the 2000s compared to the decades of the 1980s and 1990s. In this example the actual trend is increasing over this time period, so one may jump to the conclusion that this will continue with higher numbers of extreme events as we move forward in time. However, this is based on only 30 years of knowledge. The recent increase may all be part of a much longer period cycle ... perhaps the occurrence of extreme weather events peaks every 50 years with less frequent extreme weather events between the peaks. One point is that we do not have enough understanding of the climate system and natural climate variability to answer questions similar to the hypothetical question presented here. In other words, we do not have enough information to determine whether or not there has actually been any significant change in the number or intensity of extreme weather events when put into the context of the entire history of extreme weather events that have happened on Earth. It is important to realize that our lives are short compared to some of the natural cycles of climate variability. We do not experience the full range of weather and climate extremes over a lifetime. We tend to think that our notion of climate is the way things "should be" or way things "have always been." Therefore, we can easily misinterpret changes from our perspective.

Whenever a disaster occurs, people wonder what caused it to happen. Prior to the large increase in greenhouse gases due to human activity (pre 1950), we can definitely say that extreme weather events happened naturally, i.e., they were not influenced much by human activity. Today, many people are quick to attribute extreme weather events to increased greenhouse gases. Hopefully, you can see that there is no way this can be proven considering that we do not understand the natural variability in extreme events. In other words, how can we be sure that a particular extreme event would not have happened (or that it would not have been as severe) if humans did not add greenhouse gases to the atmosphere? Does this mean that increased greenhouse gases have not had a significant effect on extreme weather events? No! All we can say now is that based on our current knowledge of past climate cycles and extreme weather events, we are unable to determine if recent severe weather events are outside the range of what would occur naturally.

The idea that extreme weather events are more common and severe today and that these events will only get worse in the future because of human-added greenhouse gases is not wild speculation though. One simple line of reasoning is this ... adding greenhouse gases should cause the surface temperature of the Earth to increase, which would cause the rate of evaporation to increase. Thus, in places where it does not rain, droughts may be more severe, and since warmer air is able to hold more water vapor, in places where it does rain flooding may be more severe. These conclusions, while plausible, are far from certain. There have also been several recent studies using climate models which indicate that extreme weather events become more common and severe after greenhouse gases have been added to the atmosphere. These modeling studies do support the possibility that human-added greenhouse gases may significantly influence extreme weather events, however they cannot be used a scientific proof, since climate models are not reality.

Here is a local example about possible attribution of severe weather and climate events. We know that there have periodic severe droughts in the western United States. Evidence for this abounds, which includes tree-ring observations done right here at the University of Arizona. Based on historical evidence we can expect more severe droughts in the western United States. In other words periodic extreme droughts are normal for this region. The most severe of these droughts can be separated by hundreds of years. Thus, it is very possible that there are many people that have never experienced a severe drought, nor have their recent ancestors such as parents and grandparents. Now suppose an extreme drought does occur. People who have never experienced a drought (or do not know climate history) think this is not normal and there must be a reason this change is happening. This is the point where someone could claim that the drought was caused by humans, possibly due to greenhouse gas emissions and their subsequent influence on climate. I hope you can see why this would be difficult to prove. Of course it would also be difficult to prove that severe weather events today are not at all influenced by the higher levels of greenhouse gases in the atmosphere. There are those who say that a good reason to reduce our emissions of greenhouse gases is because this will either cause droughts in the western United States or make drought more severe. I say we should prepare for severe droughts in the western United States regardless of whether or not we can prove that it will be influenced by higher levels of greenhouse gases because we know from the past that they will happen again.

Two Perspectives on Recent Trends in Extreme Weather Events

The Intergovernmental Panel on Climate Change (IPCC) has released a set of reports, which comprise the Fifth Assessment Report (AR5) on climate change. Material from the Summary for Policymakers will be referenced in this section. The IPCC has modified some of its statements about recent changes in extreme weather events in this latest report.

In the Fourth Assessment Report (AR4) published in 2007, the following claim was made "it is likely to very likely that events such as heat waves, floods, droughts, and tropical storms have become more severe since 1960." Based on my personal research and some of the arguments presented above, I think these claims are difficult to prove given our rather poor understanding of historical climate and severe weather events and our poor understanding of the mechanisms driving past natural climate changes and fluctuations in extreme weather events. Given that these extreme weather events occur so infrequently, we would need many more years of good observations before we understand the natural variability of these events. And we need to understand the natural variation in order to determine if there has been some significant change since 1960. Here we are just considering the question as to whether or not there are provable changes in the frequency and severity of extreme weather events, without speculating on the cause of the (possible) changes.

In Table SPM.1 in the 2013 Summary to Policy Makers, the following information about the possible changes in severe weather since the year 1950 is presented. First, "medium confidence" that heat waves have increased worldwide and "likely" that heat waves have increased in Europe, Asia, and Australia. Second, "low confidence" that the intensity and duration of drought has increased. Third, "low confidence" of increases in tropical storm activity globally, but "virtually certain" in the North Atlantic since 1970. (Instructor's Note. It is interesting to me that 1970 was near the beginning of the less active part of the Atlantic multi-decadal cycle in hurricanes.) The latest AR5 has backed away from many of the claims of increased severe weather events in recent times. The previous conclusion based on the 2007 report, either implicitly or explicity stated, was that extreme weather events were becoming more severe as a result of warmer surface temperatures. Unfortunately, this connection has been accepted by the popular media and many in the public believe that there is a proven connection between higher surface temperature and extreme weather events. It is commonly presented as fact in some textbooks and TV "science" shows that extreme weather events are becoming more severe and that if the surface temperature were to increase further, we can expect even more severe weather events. You should understand that this connection is far from scientifically established and for the most part cannot be supported by available data. In this discussion we avoided the question of why surface temperature increased between 1900 and today and are just considering if there is observational evidence that proves extreme weather events are becoming more severe.

Does this rule out a connection between higher surface temperature and extreme events? The answer is no. However, what we can say is that if there is a connnection, the connection so far has been too weak for us to distinguish significant changes in extreme events from natural variations in these events.

The linked article below is from the Global Warming Policy Foundation (GWPF), which is a think tank. Their stated purpose (taken from their web page) is "to help restore balance and trust in the climate debate that is frequently distorted by prejudice and exaggeration." This group is often critical of the information and reports published by the Intergovernmental Panel on Climate Change (IPCC). You should keep that in mind when reading the essay below. Also keep in mind that this article was written to respond to some of the claims made in the 2007 AR4 on climate change. A possible increase in extreme weather events (and the subsequent toll that will have on human life and property) that may come about if the Earth's surface temperature were to increase in the future is one of the issues addressed in A Primer on the Global Death Toll from Extreme Weather Events - Context and Long Term (1900-2008)

Extreme Heat and Cold (the Silent Killers)

Perhaps because extreme heat and cold do not make spectactular news videos, many people are surprised when they hear that temperature extremes are the largest weather related cause of death. There are conflicting studies about whether heat waves or cold waves are the biggest killers, but most agree that on average, exposure to extreme temperatures kills more people each year than hurricanes, lightning, tornadoes, blizzards, etc. combined. Certainly the fact that extreme temperature does not result in much visual property damage is a major reason that deaths related to extreme temperature do not receive enough attention. Another reason is that many of the people who die from exposure are elderly, weak, or poor and mainstrean society tends to not notice these people. Another factor is related to arguments about which deaths to classify as heat or cold related. Statistics, such as "excess deaths" during extreme temperature events, which is explained below, are probably better indicators of the problem. Studies that do not consider "excess deaths", such as those reported by the US National Weather Service, may severely underestimate the number of deaths related to extreme temperature events.

The World Meterological Organization has defined a heat wave as "5 or more consecutive days where the daily high temperature is 9°F or more above average (in the warmest season)." There does not seem to be a corresponding definition for a cold wave, but I guess it would be similar. This means that the conditions necessary for a heat wave or cold wave are different for different geographical areas. For example, a week of below freezing low temperatures in Florida would consitute a cold wave, whereas in Minnesota this would cause little difficulty. Conversely, a week of high temperatures above 100°F in Chicago is a heat wave, whereas in Phoenix, this is normal.

In the popular press, more attention has been paid to dangers of heat waves in comparison to the dangers of cold waves. In recent years there have been many articles and news stories stating that exposure to extreme heat kills more people than all other weather-related causes combined, including exposure to extreme cold. However, there have been studies by reputable medical researchers that conclude the opposite ... that exposure to extreme cold is responsible for more human deaths than exposure to extreme heat. For example, the following quote is taken from an article entitled The Impact of Global Warming on Health and Mortality published in the Southern Medical Journal: "Meanwhile, it must not be forgotten that cold weather in winter causes many more deaths than heat in summer, even in most subtropical regions, and measures to control cold-related deaths need to continue." This statement is partially based on a study of heat and cold related deaths in Europe, which found that over the 10 year period from 1993-2003, "Mean annual heat related mortalities were 304 in North Finland, 445 in Athens, and 40 in London. Cold related mortalities were 2457, 2533, and 3129 respectively." The rather unbalanced coverage by the popular press makes me wonder if it is somewhat driven by alarmist fears of global warming rather than ethically presenting the whole story. For example see Extreme cold kills more people than leukemia, homicide and liver disease from the Washington Post and European Heat Wave 2003: A Global Perspective. In Extreme Weather Events Mortality and Migration the authors conclude that an important factor in the recent increase in US life expectancy is the mass migration of people from the northeastern part of the US to the warmer southwestern part of the US.

Let's just say that we are unable to conclude whether heat waves or cold waves are responsible for more human deaths. In any case, there are large differences among various studies in the number of deaths attributed to extreme heat and cold. The largest source of uncertainty is in classifying which deaths are caused by exposure to extreme temperatures. Although overall mortalilty rates tend to increase sharply during both heat and cold waves, few deaths are directly attributable to either heat stroke or hypothermia and there has been no consistent standard in reporting temperature-related mortality. The majority of deaths are ascribed to causes not commonly considered to be weather related, such as circulatory and respiratory diseases. Studies that count all of the "excess deaths" (the number of deaths above the long-term average for a region) during a heat or cold wave report significantly higher temperature-related deaths than those that only include heat stroke or hypothermia. For example, the National Weather Service statistics (shown above) will only count a death as heat or cold related if the official cause of death is reported to be due to heat stroke or hypothermia, which results in much lower estimates of temperature related deaths than studies that consider "excess deaths" during heat and cold waves. Some authors also claim that there is a lag or delay between the end of the heat or cold wave and the end of the related high mortality rates suggesting still higher death tolls related to exposure to extreme temperatures. The U.S. Centers for Disease Control consider heat and cold related death and illness to be a significant problem and have been running a public campaign to alert people to the dangers of exposure to extreme temperatures.

Both heat and cold related deaths fluctuate quite a bit from year to year. Most deaths are associated with prolonged periods of hot or cold temperatures that occur sporatically. For very severe heat and cold waves death tolls due to exposure to extreme temperatures were estimated to be as high as several or even tens of thousands, while in more moderate years few are killed by exposure to extreme temperatures. Thus, most deaths occur in a small percentage of years with extreme heat or cold waves. For example, the European heat wave in summer 2003 was particularly deadly European heatwave caused 35,000 deaths. Soon after that, the cold winters in 2005 and 2006 were estimated to have killed 29,000 and 23,000 in England and Wales alone, Why more people die in winter. This last article also points out that countries that have the mildest winters tend to have higher cold weather-related deaths than countries with very cold winters.

Perhaps the most disconcerting part of the story is that the general public seems unaware that exposure to extreme temperature kills so many people. This was pointed out with regard to heat waves in a New York Times article Most Deadly of the Natural Disasters: The Heat Wave. The article also points out that deaths caused by heat waves are typically most severe in large northern cities such as Chicago and New York because the people that live there and the buildings in which they live are not well adapted to the infrequent heat waves that strike. In southern cities high temperatures are more common, and people are more prepared to deal with the heat. Many public health officials believe that a large fraction of the excess deaths that occur during heat and cold waves are avoidable through public education and social services.

Although many claim that there have been recent increases in heat waves and heat related deaths due to climate change, to my knowledge there is no scientific consensus indicating that the frequency or severity of heat waves has increased significantly in recent years. Some studies conclude that there have been no changes in the frequency or intensity of heat waves, while the 2013 IPCC AR5 Report claims that it is "likely" that the frequency of heat waves has increased in parts of Europe, Asia, and Australia since 1950. Some will then argue that human activity is causing global warming, therefore, there will be more heat waves, therefore we must take immediate action to reduce emissions of greenhouse gases. This is often presented as a fact when in reality the connection between global warming and the possible change in the frequency and severity of heat and cold waves is highly uncertain. To be fair, those who claim that many more people may die in heat waves with global warming, should also be saying that many fewer people may die due to exposure to extreme cold. Regardless, deaths could be more effectively reduced by making sure people have proper shelter and medical aid during heat and cold waves rather than worrying about whether or not global warming may be making it slightly worse.

[Home] [Lectures] [Next]